![]() |
SIMULATIONS FOR ENGINEERING EDUCATION |
![]() |
|
|
10 DOF Robot Completing DOE D&D Simulation
This is a simulation of a 10 DOF Manipulator completing a set of tasks associated with DOE Decontamination & Dismantlement (D&D). Although redundant systems have numerous advantages over their non-redundant counterparts (work around obstacles, larger workspaces, continued operation even if a joint fails, better performance, etc), they are rarely deployed due to the increased complexity associated with their operation. As seen in the the earlier self-motion example, an infinite number of solutions is possible, but determining the best (of even a valid) solution is not simple for an operator who does not have an advanced degree in robotics. The high level objective of the Redundancy Resolution Strategy (RRS) demonstrated here is long-term efficient operation by a non-expert user. This objective is attained by monitoring critical boundaries that are associated with systems constraints (joint travel limits, torque limits, collisions, kinematic singularities, etc.) and then using performance criteria to help the system find configurations that move the manipulator away from the constraints whenever the critical boundary is crossed. When no critical boundaries have been crossed, the system considers only criteria associated with efficiency (minimize joint velocities, joint torques, kinetic energy, power flow, etc.)
The colored bulbs on the base of the manipulator let you know which critical boundaries have been crossed during the simulation. In the simulation, the robot completes multiple tasks including tool exchanges, cutting, pick-and-place, and using a pneumatic wrench. Every motion is completed in EE space. The base location of the manipulator as well as the location and orientation of the collected bricks were randomly generated. The entire simulation (over 38,000 set points) was completed without any user intervention required.
Instructions:
Contact Mitch Pryor at mpryor@mail.utexas.edu if you would like more information on this research.
|
![]()
Comments about this page should be directed to mtisius@mail.utexas.edu
© 2001 Robotics Research Group